We are proposing a five-year study of adults with major depressive disorder refractory to treatment with fluoxetine. We want to address the basic clinical question: once a fluoxetine trial of adequate duration and dosage has failed, what should be tried next? Should one increase the dosage of fluoxetine, or maintain the same dosage and add either low-dose of a tricylic antidepressant or lithium? The scope of our study is to compare the efficacy of these three different treatment strategies in this area in great need of controlled research. The critical clinical issue is whether the use of adjunctive medications offers any additional benefit given the popularity and increasing use of this strategy by clinicians. An additional aim of the study is to assess the safety and tolerability of these treatments. The study involves the enrollment over 48 months of 500 patients with major depression who will be treated openly with fluoxetine 20 mg/day for 8 weeks. The subjects will be drawn from an outpatient sample of patients with major depressive disorder, diagnosed by the use of the Structured Clinical Interview for DSM-III-R - Patient Edition (Spitzer et al, 1989), and with an initial 17-item Hamilton Rating Scale for Depression (HAM-D-17) (Hamilton, 1960) score greater than 16. If these depressed patients fail to respond to this 8-week open treatment with fluoxetine, they then will become eligible for a four-week, three-arm, double-blind, triple-dummy phase of the study where they will be randomized to one of three different treatments: fluoxetine 40-60 mg/day, fluoxetine 20 mg and desipramine 25-50 mg/day, and fluoxetine 20 mg and lithium 300-600 mg/day. Refractoriness to treatment will be defined as a failure to achieve a 50% or greater reduction in the HAM-D-17 scores and a HAM-D-17 score greater than 10 at the time of enrollment. There will be a prospective stratification of the patients for the randomization based on the fact that they will be either partial responders or non responders: partial responders will be considered those patients who had experienced a > 25% and < 50% reduction in their HAM-D-17 score or have a HAM-D-17 score or have a HAM-D-17 score < 16; non responders will be considered those patients who have had a less than 25% reduction in their HAM-D-17 score from the beginning of treatment and have a HAM-D-17 score greater than 16. The study extends our previous research on drug therapy strategies for treatment-refractory depression and examines the efficacy of three different treatments in a depressed population whose refractoriness to the same antidepressant has been defined prospectively rather than historically or retrospectively.